The paper presents experiments and analysis of energy confinement on the CTX spheromak. Compared to previous published results from 0.4 m radius flux conservers, in a 0.67 m radius mesh flux conserver (with the current density kept constant), the magnetic field increases while the plasma density is kept the same. However, the electron temperature does not rise, and hence (fi) yol drops. The plasma resistivity remains constant (the resistance drops as the size increases), and the energy confinement time stays the same. Plasma energy content results from spheromaks during sustainment by helicity injection are also presented and show confinement equivalent to that during the decay phase. Increased magnetic field in the same size experiment produces very little improvement in electron temperature and a decrease in confinement time. The resistive decay time is found to be empirically independent of the core electron temperature. It is, instead, proportional to the strength of the magnetic field at constant plasma density, while the ratio of magnetic field to decay time depends on plasma density, consistently with ionization breakdown at the edge of the spheromak dominating helicity dissipation. The possible causes of this observed confinement are examined separately in detailed quantitative and qualitative studies. Absolutely calibrated multichord bolometry shows that impurity radiation is not the cause of the low electron temperatures. The particle confinement time has increased with size, but does not show an increase with increasing magnetic field. At the lower (/3) vo) of the larger experiment, the particle replacement power cannot explain the unaccounted energy losses. Any important role of pressure driven modes in the CTX energy balance is shown to be inconsistent with the available CTX data. The possibility that rotating coherent current driven kink modes can seriously degrade energy confinement is evaluated and discounted owing to the lack of improvement when the modes are absent. The role of anomalous ion heating is examined, and the available data are presented. Finally, a hypothesis explaining these results is presented, suggestions for future work are made, and the results are summarized.